A Methodology for Comparing and Evaluating Seriation Algorithms Applied to Archaeological Data

Summary

This is an abstract from the "SAA 2019: General Sessions" session, at the 84th annual meeting of the Society for American Archaeology.

According to recent literature, correspondence analysis is the method of choice for frequency seriation. However, this does not consider the effects of data heterogeneity or typology on the orderings produced by this method. This relates to a more fundamental issue of how to evaluate the effects of heterogeneity and typology on seriation results, as well as how to determine which of a set of seriation algorithms produces the more likely seriation ordering on a particular data set, and if so, why? In this paper we present a new methodological framework that: (1) identifies which parts of a data set are amenable to seriation, (2) identifies the likely number of minimally heterogeneous components in a data set in a way that is sensitive to emic distinctions, and (3) operationalizes a Bayesian stochastic seriation model to evaluate which of a set of seriation algorithms produces the more likely ordering on a given data set. When this new framework is applied to heterogeneous mortuary data set from a coastal Chumash village in southern California, the results suggest that correspondence analysis does not produce a likely seriation ordering.

This Resource is Part of the Following Collections

• Society for American Archaeology 84th Annual Meeting, Albuquerque, NM (2019)

Keywords

Spatial Coverage

Record Identifiers

Abstract Id(s): 23805